From: Yao Qi <yao@codesourcery.com>
To: Pedro Alves <palves@redhat.com>
Cc: Joel Brobecker <brobecker@adacore.com>,
Eli Zaretskii <eliz@gnu.org>, <gdb-patches@sourceware.org>
Subject: Re: [PATCH] Unbuffer stdout and stderr on windows
Date: Wed, 28 Aug 2013 07:23:00 -0000 [thread overview]
Message-ID: <521DA52C.9020003@codesourcery.com> (raw)
In-Reply-To: <521D0E8A.80303@redhat.com>
On 08/28/2013 04:39 AM, Pedro Alves wrote:
> Better say that "it is not used on Windows hosts". But I think
> it'd be better even to say:
>
> /* Method 'to_write_async_safe' is not overwritten, because
> there's no way to flushing a stream in an async-safe manner.
> Fortunately, it doesn't really matter, because:
> - that method is only used for printing internal debug output
> from signal handlers.
> - Windows hosts don't have a concept of async-safeness. Signal
> handlers run in a separate thread, so they can call
> the regular non-async-safe output routines freely. */
OK.
[...]
> hide that in ui-file.c itself. So e.g., we'd have a new stderr_fileopen
> method, and then both places that create gdb_stderr would use it, like:
>
> -gdb_stderr = stdio_fileopen (stderr);
> +gdb_stderr = stderr_fileopen ();
>
> We could also that whole set of Windows-specific comments there
> too.
Right. Function stderr_fileopen is added in the updated patch.
What do you think?
--
Yao (é½å°§)
gdb:
2013-08-28 Yao Qi <yao@codesourcery.com>
Pedro Alves <palves@redhat.com>
* event-top.c (gdb_setup_readline): Call stderr_fileopen
instead of stdio_fileopen.
* main.c (captured_main) [__MINGW32__]: Set stderr unbuffered.
Call stderr_fileopen instead of stdio_fileopen.
* ui-file.c (stderr_file_write): New function.
(stderr_file_fputs): New function.
(stderr_fileopen): New function.
* ui-file.h (stderr_fileopen): Declare.
---
gdb/event-top.c | 2 +-
gdb/main.c | 9 ++++++++-
gdb/ui-file.c | 54 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
gdb/ui-file.h | 3 +++
4 files changed, 66 insertions(+), 2 deletions(-)
diff --git a/gdb/event-top.c b/gdb/event-top.c
index f00ab7d..f1d55b3 100644
--- a/gdb/event-top.c
+++ b/gdb/event-top.c
@@ -955,7 +955,7 @@ gdb_setup_readline (void)
time. */
if (!batch_silent)
gdb_stdout = stdio_fileopen (stdout);
- gdb_stderr = stdio_fileopen (stderr);
+ gdb_stderr = stderr_fileopen ();
gdb_stdlog = gdb_stderr; /* for moment */
gdb_stdtarg = gdb_stderr; /* for moment */
gdb_stdtargerr = gdb_stderr; /* for moment */
diff --git a/gdb/main.c b/gdb/main.c
index 1c240e4..11f4b03 100644
--- a/gdb/main.c
+++ b/gdb/main.c
@@ -375,8 +375,15 @@ captured_main (void *data)
saved_command_line[0] = '\0';
instream = stdin;
+#ifdef __MINGW32__
+ /* Ensure stderr is unbuffered. A Cygwin pty or pipe is implemented
+ as a Windows pipe, and Windows buffers on pipes. */
+ setvbuf (stderr, NULL, _IONBF, BUFSIZ);
+#endif
+
gdb_stdout = stdio_fileopen (stdout);
- gdb_stderr = stdio_fileopen (stderr);
+ gdb_stderr = stderr_fileopen ();
+
gdb_stdlog = gdb_stderr; /* for moment */
gdb_stdtarg = gdb_stderr; /* for moment */
gdb_stdin = stdio_fileopen (stdin);
diff --git a/gdb/ui-file.c b/gdb/ui-file.c
index cf5a86d..d29b8da 100644
--- a/gdb/ui-file.c
+++ b/gdb/ui-file.c
@@ -654,6 +654,60 @@ stdio_file_fseek (struct ui_file *file, long offset, int whence)
return fseek (stdio->file, offset, whence);
}
+/* This is the implementation of ui_file method to_write for stderr.
+ gdb_stdout is flushed before writing to gdb_stderr. */
+
+static void
+stderr_file_write (struct ui_file *file, const char *buf, long length_buf)
+{
+ gdb_flush (gdb_stdout);
+ stdio_file_write (file, buf, length_buf);
+}
+
+/* This is the implementation of ui_file method to_fputs for stderr.
+ gdb_stdout is flushed before writing to gdb_stderr. */
+
+static void
+stderr_file_fputs (const char *linebuffer, struct ui_file *file)
+{
+ gdb_flush (gdb_stdout);
+ stdio_file_fputs (linebuffer, file);
+}
+
+/* Create a ui_file from stderr. */
+
+struct ui_file *
+stderr_fileopen (void)
+{
+ struct ui_file *ui_file = stdio_fileopen (stderr);
+
+#ifdef __MINGW32__
+ /* There is no real line-buffering on Windows, see
+ http://msdn.microsoft.com/en-us/library/86cebhfs%28v=vs.71%29.aspx
+ so the stdout is either fully-buffered or non-buffered. We can't
+ make stdout non-buffered, because of two concerns,
+ 1. non-buffering hurts performance,
+ 2. non-buffering may change GDB's behavior when it is interacting
+ with front-end, such as Emacs.
+
+ We decided to leave stdout as fully buffered, but flush it first
+ when something is written to stderr. */
+
+ /* Method 'to_write_async_safe' is not overwritten, because there's
+ no way to flushing a stream in an async-safe manner.
+ Fortunately, it doesn't really matter, because:
+ - that method is only used for printing internal debug output
+ from signal handlers.
+ - Windows hosts don't have a concept of async-safeness. Signal
+ handlers run in a separate thread, so they can call
+ the regular non-async-safe output routines freely. */
+ set_ui_file_write (ui_file, stderr_file_write);
+ set_ui_file_fputs (ui_file, stderr_file_fputs);
+#endif
+
+ return ui_file;
+}
+
/* Like fdopen(). Create a ui_file from a previously opened FILE. */
struct ui_file *
diff --git a/gdb/ui-file.h b/gdb/ui-file.h
index 9fef68c..dca800f 100644
--- a/gdb/ui-file.h
+++ b/gdb/ui-file.h
@@ -129,6 +129,9 @@ extern struct ui_file *mem_fileopen (void);
/* Open/create a STDIO based UI_FILE using the already open FILE. */
extern struct ui_file *stdio_fileopen (FILE *file);
+extern struct ui_file *stderr_fileopen (void);
+
+
/* Open NAME returning an STDIO based UI_FILE. */
extern struct ui_file *gdb_fopen (const char *name, const char *mode);
--
1.7.7.6
next prev parent reply other threads:[~2013-08-28 7:23 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-07-22 3:07 Yao Qi
2013-07-22 15:41 ` Eli Zaretskii
2013-07-23 6:35 ` Yao Qi
2013-07-23 17:52 ` Eli Zaretskii
2013-07-29 19:26 ` Christopher Faylor
2013-07-29 19:30 ` Eli Zaretskii
2013-07-29 19:51 ` Pedro Alves
2013-07-31 3:40 ` Christopher Faylor
2013-08-12 21:11 ` Joel Brobecker
2013-08-13 17:28 ` Christopher Faylor
2013-08-13 18:08 ` Eli Zaretskii
2013-08-14 0:05 ` Joel Brobecker
2013-08-15 17:36 ` Christopher Faylor
2013-08-15 17:44 ` Eli Zaretskii
2013-08-15 17:59 ` Christopher Faylor
2013-08-15 18:44 ` Eli Zaretskii
2013-08-16 11:46 ` Pedro Alves
2013-08-16 12:34 ` Yao Qi
2013-08-16 13:20 ` Eli Zaretskii
2013-08-16 13:37 ` Pedro Alves
2013-08-16 14:03 ` Eli Zaretskii
2013-08-16 14:21 ` Pedro Alves
2013-08-16 14:57 ` Eli Zaretskii
2013-08-16 15:10 ` Pedro Alves
2013-08-16 15:24 ` Pedro Alves
2013-08-16 15:43 ` Eli Zaretskii
2013-08-16 16:41 ` Christopher Faylor
2013-08-16 15:41 ` Eli Zaretskii
2013-08-22 6:14 ` Yao Qi
2013-08-22 14:18 ` Joel Brobecker
2013-08-23 2:20 ` Yao Qi
2013-08-23 13:38 ` Joel Brobecker
2013-08-27 20:39 ` Pedro Alves
2013-08-28 7:23 ` Yao Qi [this message]
2013-08-28 9:39 ` Pedro Alves
2013-08-28 12:25 ` Yao Qi
2013-08-16 13:17 ` Eli Zaretskii
2013-08-16 13:30 ` Pedro Alves
2013-08-16 13:42 ` Eli Zaretskii
2013-08-16 14:13 ` Pedro Alves
2013-08-16 14:44 ` Eli Zaretskii
2013-08-16 15:05 ` Pedro Alves
2013-08-16 15:13 ` Eli Zaretskii
2013-07-29 19:30 ` Christopher Faylor
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=521DA52C.9020003@codesourcery.com \
--to=yao@codesourcery.com \
--cc=brobecker@adacore.com \
--cc=eliz@gnu.org \
--cc=gdb-patches@sourceware.org \
--cc=palves@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox